• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

SCVMM 2012 R2 networking help

goffries

n00b
Joined
Aug 17, 2005
Messages
21
I was recently thrown into a virtualization project due to me knowing more than nothing about the process, but the environment they want created is fairly large and I was hoping someone here could give me some advice.

I am deploying a new VMM 2012 R2 server with 44 2012 r2 hyper-v hosts, each have a quad port nic and 2 onboard nics, all ports for all hosts are on the same vlan and subnet. I have been reading some blogs about SCVMM best practices and one of the primary things it says is try to separate out different types of traffic using VMM. I'm afraid I'm a bit of a dud when it comes to networking and i was hoping someone could help simplify the process.

If all the hosts are on the same vlan and subnet how can you separate out the live migration traffic from the host management and guest traffic?

I do not have any control over the switches at my company and I would like to be able to avoid any direct manipulation of them.

If any additional info is needed i would be happy to provide.
 
woah.. what?
You're going to deploy 44 host servers? How many VM's are you talking?
Youre going to go that large and only put a single quad port 1gb/sec nic in them? At that level you need to step up to 10GbE links.

You've not given near enough information to really help you. Are you doing any iSCSI?
Are the hyper-v hosts going to be clustered in any way? If not I'm not sure why you're asking about live migration.
 
4-8 guests per host. We are not doing any Iscsi, all shared storage will be provided via fiber. All hosts are in 5 node clusters. Unfortunately our switches are all 1GbE with 10GbE uplinks. So 10GbE per host is not feasible right now.

I was hoping to do something like 1 of the onboard nics dedicated to host management, 1 to live migration, and then team the quad port nic for regular VM traffic.
 
You will want redundancy at the MGMT and Live Motion layer, however you are missing a couple of things here. If you are using Hyper-V in a Failover Cluster you will need a dedicated heartbeat network as well and if you will be using Cluster Shared Volumes, which I would recommend, you will need dedicated network for that as well.

I would get another quad port nic and make sure all networks have redundancy at the very least.

Failover Clustering with Hyper-V has strict requirements and they should be followed to avoid issues.

When I first rolled out Hyper-V Failover Cluster in 2010, I didn't have a dedicated Cluster Heartbeat and i can't tell you the issues.

I researched like mad, and went over all best practice and recommendations from MS experts. I set it up with all required networks and everything worked and is still working like a charm at my old job.

Of course this is with Hyper-V with Failover cluster on 2k8 R2..so i'll defer to others with 2012 experience.
 
One option from a networking perspective is to create one Hyper-V network team with 1 port from the onboard and 3 of the ports from the quad NIC then build a vSwitch on top of that. Then you create virtual adapters for your host management, cluster heartbeats, and VM network adapters off that vSwitch. Assign VLANs and QoS to the virtual adapters as you see fit.

Do not avoid configuring the switches. You're building a large environment; why do it wrong? You should be defining separate VLANs and subnets for host management, Live Migration, cluster heartbeats, VM network adapters, etc.. Do not put everything on the same subnet and VLAN.

The remaining two NICs can be used for Live Migration. Create an active/standby network team with them and give the new team adapter your Live Migration IP. If your NICs support RDMA, you can skip the network team for Live Migration and simply assign an IP to both adapters on the same subnet/VLAN and use SMB for Live Migration and get double the bandwidth.

Now you'll have proper traffic segregation and failover plus the ability to configure QoS in Hyper-V for your host and VM traffic.
 
Back
Top